t.t'r¡
ALTERNAÎIVE MODELS OF lHE SOTJND FIEIÐ
IN
A REVERBERAI{T ROOMby
Thomas
J. Munro
B. Sc.A thesÍs
presentedto the
Faeultyof
Engineeringof the Unfverslty of
Adelaldefor the
Degreeof
Masterof
Englneering SclenceDepartment
of
Mechanfcal EnglneeringUniversity of
AdelaideJanuary 1982
TABLE OF CONTENTS
Summary
Statement
of
0ri gì naì ity
Acknowl edgements
chapter
I
AC0USTIC MEASUREMENTS IN REVERBERATI0N R00MSl.l
Introduct'ion1.2
tory of
Reverberant Sound Measurements1'l
1V
vi
vìi
I I 3
4 9 12
Ear'ly ̡lork
'in
Room AcousticsThe
Statistical Variation of
MeasuredQuantì t'ies
Determ'i nat'ion
of
Sound Power Iniecti
on i nReverberant F'iel ds
Methods
of
Obtainingan Accurate Space Averageof
l4easured Quantì b'iesStati
sti cal
Vari a'b'ion ìn
Decayi ng Sound Fi el ds Comparisonof Stat'ist'ical
and Modal Modelsof
Room Behavi our
Predict'ion
of
Rateof
Decay byAìternative
Model s
1.3
Standard l4easurenent Methods Power LevelAbsorption Coeff
ici
ent Transmi ss'ion Loss ndnd nd
His 1.2 1.2
| .2.5
1 .2,'o
1.3.1
r.3.2
I .3.3
2
1 .2.3
1 .2.4
1 .2.7
l4 l5
Sou Sou Sou
17
l8
20 21 25 27
Chapter 2 2.1
2.6 Chapter 3
3.1 3.2 2.2 2.3
ALTERNATIVE THEORETICAL DEVELOPMENTS I ntroduct i on
The hlave Equation and Soìutions
Steady
State
Behaviourof the
Modal SoundField
2 .3
.1
Equal Energy Di s'lri buti on 2.3.2
Equaì Power F I owModal Descri
pt'ion of the
Decayi ng Sound F'iel dModal Damping by
the
Acoust'ic Ray Approxitnalion2.5.1
Prediction of
Modaì Decay Rate2.5.2 Norris-Eyrinq
Equation asa
Spec'ial Case M'iscel I aneous Formul asMATERIALS AND METHODS
Phys'icaì Fac'i I
it'ies
Instrumentat'ion, Sì gnaì Generat'ion and Measurement
3.2.1
Generationof
Sound3.2.2
Measurement and Process'ingof
Sounci PressureLevel s
Experìmental Methods
3.3..l
Measurementof
Modal Decay Rate3.3.2
Measurementof
SteadyState
Modal Energy Level37 30 30 32
37 40
41
43 43 48
5l 5l
53 54 56 57 57 60
?.4 2.5
49
3.3
Chapter 4 4.1
4.2 4.3
4.4 Chapter 5
5.1 5.2 Chapter 6 References
4.3. I
4.3.2
RESULTS: MODAL BEHAVIOUR IN THE EMPTY ROOM
Veri f i cati on
of
l4odal Behav'iour4.1.1
Modal Behaviourin
SteadyState
Conditions4.1
.2
Modal Behaviour
ìn a
Decay'i ng Sound Fi el dThe
llature of
Moda'l DecaYAnaìysì
s of the
Modal Rateof
DecaY4.3.3
The
Tradìtional
ltlodal Approach Modal Damping bythe
RaY Tracing Approxi rnatì onProbable Reasons
for the Fa'ilure of the
Modeìsto
Pred'ictActual
BehaviourInitial
EnergyLevel of
Room ModesTHE ROTATING DIFFUSER, ITS EFFECT ON MODAL STRUCTURE
Previous l,Jork
lvith
Rotat'ingDiffusers
Experimental ResultsCONCLUS IONS
lll
Page
62 62 62 63 68 73 77
79 83 84
90 9t
92
l0l
103
]V
SUMMARY
Current standard test
methodsfor acoustic
measurementsnade
inreverberant
roomshave
beenfound to y'ieìd inconsìstent results at
lolvfrequencies. The theoret'ical basis of the standard test
methodsrelìes
upon
the
assumptionthat the reverberant
soundfield can be
modeìed ascompìeteìy d'iffuse ab
alI
frequencies. Thus,
'inconsistenc'ies have
beenattributed to a lack of diffusivity at low frequencies. Th'is
thesisexami
nes the
appropriateness of the
diffuse fi
eld
assumption at
I owfrequencìes and
showsthat the strong modal characteristics of a
lowfrequency reverberant
soundfield precìudes the possibi'lìty of
mode'lingsuch a
field
asdìffuse.
A
nlodelof the reverberant
soundfield which includes the
observedmodal characteri
st'ics and
is
based upon welI
known solut'ions to the
waveequa'ti
on
is
presented.
Th js
model 'is
dependenton the
accurate predicti
onof the steady state
ampì'itude
andthe
decayrate of
i nd'ividual
modes.
I npred'icti
ng the steady state
ampì'itude of the
modes contribui'ing to
areverberant
soundf
ield, the
assumpt'ion is
usually
madee'ither
i mpì ic'itìy or expììcìt1y that there 'is equal
energydistributioç
between modes. In th'is
thes'is, an
alternati ve
approach basedon the
assumption that
there'is
equal powerflow to alì
modesvlill also
be consìdered. These assumpt'ionsare
di scussed andthe'ir
i nf I uenceon the
rel ative
diff
erence 'i nthe
steadystate
soundpressure
Ievel of
any two modes is
analyti ca'lly
'i nvest'igated.Two models
for predicting the decay rate of ind'iv'idual
modes arealso
examined. Thefjrst is
based uponan
approx'imatesolution to the
waveequation for a very ìightìy
damped room. The secondis
based upona
com-b'ina'li
on of the modal
approachand the ray tracì ng
approachas
appì 'iedto the rectangu'lar room. Both
modelsreìy
uponthe
assumptionthat
thewal I
s can be accurateìy
modeled as
'loca'lly
reactive and that the
wal lV
.impedance can be modeled
as constant over the wall
surfaces and throughoutthe
frequency rangeinvestigated.
Experintental results are presented for
comparisonof each
model.Results ind'icate that the wall surfaces of the
room cannot be modeled asbeing locally
reac'biveas theìr
responseto excitation by the
reverberant soundfjeld in the
frequencyrange'investjgated is characterized by
modaìvibration. In
consequenceìt
js
shownthat neìther model
accurate'lypred'icts the relative rates of
decayof the
measured modes. Therelative
ampì i
tudes of
i nd'ividual
modesare seen to be h'igh'ly
dependenton
theeffective coupìing
between each mode andthe
soundsource.
Th'isis
shownto
behighly
unpredictablefor
commonly used sound sources.A
wideìy
recommended meansof
enhanc'ingthe
diff
usìon
'in a
rever-berant room is the installation of a rotating diffuser. Th'is
thes'isexamines
the effects of a rotatìng diffuser on the
modalcharacteristics
of a reverberant
soundfield
and showsthat jt substantially disrupts
the nodalstructure of the
soundfield.
vl
To
the best of the candidate's
knowledge andbelief'
thi s
thesìs
contains no
material
which has
beenaccepted
for the
awardof
any degreeor
dipìomain
anyUnì versi
ty, and
conta'ins no
material
previ ouslypubìished or written by another person, except
where due reference'is
made'in the text.
Thomas
J.
Munrovii
ACKNOt.lLEDGEMENTS
The
work
describedin this thesis
wascarried out in the
Departntentof
Mechanical
Engi neerìng of the Un'ivers'ity of
Adela'ide
underthe
aeg'i sof
Professor
R . E. Luxton.
Theautholi s
'i ndebtedto
Professor Luxton f orthe opportunity to carry out th'is
research.To Dr
D .A.
B'i es, who
supervised the
work, the author
is
deepìy'indebted, both for his constant interest
and encouragement throughout the courseof the
research andfor his helpful
suggest'ionsduring the
prepara-ti
onof
th'is
thesi s .To the University staff, thanks are
duefor their assistance jn
thepreparati
on of the
f aci I'iti es
in
wh'ichthe work was
compìeted, and
i npart'icular to
MrH.
Bode and MrR.
Curt'in.The author gratefully
acknowledgesthe Australian
American Eciuca-ti onal
Foundation who
provided the
f inanc'ial support wh'ich
made the conrpletionof this
work possible.